Differentiation of deep venous thrombosis from femoral vein mixing artifact on routine abdominopelvic CT

CT IVC venogram: normalized quantitative criteria for patency and thrombosis

PURPOSE

Establish normal attenuation ratios for vein to artery on CT IVC venogram and determine a vascular attenuation ratio diagnostic of thrombus.

METHODS

This retrospective, HIPAA-compliant study included 56 CT IVC venograms. Images were reviewed for the presence of femoral vein or IVC thrombus. Attenuation ratios for each vein and its corresponding artery were calculated by two observers and averaged in four venous stations (right and left femoral veins, and IVC at the confluence of the iliac veins and at the left renal vein). The reference standard for the absence of thrombus was clinical follow-up and for the presence of thrombus it was thrombectomy or catheter venogram. Receiver operating characteristic (ROC) analysis was performed using ratios from one venous station and threshold for thrombus was determined using the Youden's index.

RESULTS

36 of 56 CTs demonstrated no thrombus. 20 CTs demonstrated thrombus, confirmed in eight patients. For CTs with no thrombus, median ratios among the venous stations ranged from 0.89 (IQR 0.83-0.93) to 0.91 (IQR 0.86-0.94). ROC analysis of ratios from a single representative station (left femoral vein, n = 4 confirmed clots) demonstrated an area under the curve (AUC) of 0.994 (p = 0.001) and a threshold of 0.67 for diagnosing thrombus [sensitivity 100% (95% CI 39.76-100%), specificity 97.5% (86.84-99.94%)].

CONCLUSION

The normal attenuation ratio of vein to artery in the absence of venous thrombus on a 3-min delay CT IVC venogram is approximately 0.91. A ratio less than 0.67 is highly suggestive of thrombus.